Ain ett » anling ~t~ tx > Oe nS oe =! - = taf? Se hate Fr “7 Gee “2s
ee ee I aries Sree ears ie a BE LT LL ETT IAS STO CR LE BLE DPE I SHO
Peete ere < SBS Pele Ao ot ana Sonal ee St Sire aS ~ Fa 2
a S 4 Be Perens RA Ate d I fa linte #0 PIA owe Cegape en ore )
7 4 = A Ret hi Fale! > ak = of
orate are ihe rat as a
AE Hy Car te Or ete Se
Se eee
Sepa
v. Oe *$. SSS
eS re a
Pte ree ee
“ * Ay ra tte ir
eet, a RIE T a a ean bs
ers
ade Enh e Stee yaar
Fon we Paka og of: a aie 5 ort, ata fs
cptabrrn pire RS ge i ale Tike Pinte Spy epee ys vere
eo } tio j gaan
see ea rece beef fa Cppadisen eek Kinet abd wip ape
Me Ket pret >
sth : i ee! = ane Whe Belo ar pete’: )
ns gs Pas ae Na My ie oth Bet Stew San ae inte i = et yh Thm he Mest date ‘ht Rhee Be IA wt oP om PA
i SOAS Artes ay a. Yreka Rec} Ae eS ae ae ti eet ee ate he! ra =e m - 4
Sire si ti Seen ass Sete recut aie i Soe tet SEES :
Maha hoh Leeann’ Tos Se tote PRs . os et = x vA aati aR te
* * ‘2 pe eee ®t ewer hs, Paar Oey Faerie gg
ta Me eB ce ote aM ote Beh BoM eke BN, BT Satin Sli Me Ee ™
SITES a hs cs eet a a aoe Tat ea Ney Leena athe atime
aN ata Ml MM Mbp Fae ms Be awh Total 4 Pere wr te fe ea a De ta
Th Gee RIC ERK A RR oR Eien ema ge aS
~ oi SS zi >. oie = eo iy? 2
ed, ane Rae Mette athe hee hin! . cape nw erate rye iret ee Ce Fah Ra tet I el
BP Be Rats HOM La Ne ag he Babety MMe te MUR Rela SUAS S 6: Rete NA-Setere teat te eee ee Prana Aa hy!
an Nm oe Sa tm te GE ee he Oe he Nee eh 3h Paras MANA Atel Sint rete eh Soe eh ~
ta -Sathade Sea hs"t 0.» Sea ¥aWeehe*- 4= 8d We Re Ten ee Re, Beers ee Men, Setarn He Sn Se Nahe wz =A - take Fe
that Ratha tine > SATIS IN ee he cp Mm bie i Mn PA then Ma Mar tm Su hy Ae Komik Poe eH ta Ms en Tin Me oe Sa,
ee ee en ee ee
UA ‘. Sq Q SA 8 Mathie Boh. MedE DD 6
Fs Sere oe Pare
es
ree a te a
ee Ne ee re ee ae
etn Ha hig Pathe g Hn Tha Dame Ripe Datel ge tne
Le ee A OR LESS A
Pt Let Ooi otek D
Pe ee
a Rete rea in ee er
Tete te es eincg ad Saye
ma tS ee te TE
FF eats eM ated ee, Das
Tallon Win Ma Tn "on thes tae Tat mtn Rates tet
TS thathe 8 Bethy THF he tae %
ata Fy
5 tate Moc <thneat
Fea SRS
Tea pea cin tie eas tno ehh
Teh State
eae:
Vol. 35/No. 3
August 1968
Public Roads
A JOURNAL OF HIGHWAY RESEARCH
U.S. DEPARTMENT OF TRANSPORTATION
FEDERAL HIGHWAY ADMINISTRATION
BUREAU OF PUBLIC ROADS
Public Roads
A JOURNAL OF HIGHWAY RESEARCH
U.S. DEPARTMENT OF TRANSPORTATION ©
Published Bimonthly ALAN S. BOYD, Secretary
Harry C. Secrest, Managing Editor @ Fran Faulkner, Editor FEDERAL HIGHWAY ADMINISTRATION
Joan H. Kinbar, Assistant Editor LOWELL K. BRIDWELL, Administrator
BUREAU OF PUBLIC ROADS
August 1968 / Vol. 35, No. 3 F. C. TURNER, Director
THE BUREAU OF PUBLIC ROADS
FEDERAL HIGHWAY ADMINISTRATION
U.S. DEPARTMENT OF TRANSPORTATION
Washington, D.C. 20591
FHWA REGIONAL OFFICES
No. 1. 4 Normanskill Blvd., Delmar, N.Y. 12054.
CONTENTS Connecticut, Maine, Massachusetts, New
Hampshire, New Jersey, New York, Rhode
Jsland, Vermont, and Puerto Rico.
, ~ ; No. 2. 1633 Federal Building, 31 Hopkins
Passing Aid System I, Initial Experiments, Place Baltimore. Md. Boge e
by Duke Niebur .........-. 0000. 61 Delaware, District of Columbia, Maryland,
Ohio, Pennsylvania, Virginia, and West Vir-
inia.
Interstate System Accident Research—Study II, No: 4 1720 Peachtree Rd., N.W., Atlanta, Ga.
Interim Report II, by Julie Anna Cirillo... . 71 30309.
Alabama, Florida, Georgia, Mississippi, North
Carolina, South Carolina, and Tennessee.
New: Pablicationvn. aa.) <- fee 2 ee ee 76 No. 4. 18209 Dixie Highway, Homewood, III.
60430.
Iilinois, Indiana, Kentucky, Michigan, and Wis-
consin.
No. 5. Civic Center Station, Kansas City, Mo.
64106.
lowa, Kansas, Minnesota, Missouri, Nebraska,
North Dakota, and South Dakota.
No. 6. 819 Taylor St., Fort Worth, Tex. 76102.
Arkansas, Louisiana, Oklahoma and Texas.
No. 7. 450 Golden Gate Ave., Box 36096, San
Francisco, Calif. 94102.
Arizona, California, Hawaii, and Nevada. .
No. 8. 412 Mohawk Bldg., 222 SW. Morrison _
St., Portland, Oreg. 97204. ,
Alaska, Idaho, Montana, Oregon, and
Washington.
No. 9. Denver Federal Center, Bidg. 40, Den-
ver, Colo. 80225.
Colorado, New Mexico, Utah, and Wyoming.
No. 15. 1000 N. Glebe Rd., Arlington, Va.
22201.
Eastern Federal Highway Projects
No. 19. Apartado Q, San Jose, Costa Rica.
Inter-American Highway: Costa Rica, Guate-
mala, Nicaragua, and Panama.
COVER
Public Roads, A Journal of Highway Research, is sold
by the Superintendent of Documents, Government
Tonite kant "y GRE Printing Office, Washington, D.C. 20402, at $1.50 per
Trathe backed adap Maryland year (50 cents additional for foreign mailing) or 25
highway U.S. 301 before it be- cents per single copy. Subscriptions are available for
came a dual highway. Public 1-, 2-, or 3-year periods. Free distribution is limited to
public officials actually engaged in planning or con-
Roads research is endeavoring structing highways and to instructors of highway engi-
to solve the problem of passing neering. There are no vacancies in the free list at
vehicles E i , present. ;
é ; wd ond lane highways. Use of funds for printing this publication has been ap-
See article beginning on op- proved by the Director of the Bureau of the Budget,
posite page. March 16, 1966.
SE ee Se a ee et ed
Contents of this publication may be reprinted.
Mention of source is requested.
Vhen deciding whether to pass or not to
ass on 2-lane highways, motorists in the
uture may be assisted by electronic
ystems.
Passing Aid System |
Initial Experiments
Reported by DUKE NIEBUR, Highway Research Engineer,
Economics and Requirements Division
Development of a traffic system to aid motorists in passing vehicles on 2-lane
rural highways is one of the chief objectives of the Public Roads research and
development program. Anyone who has driven on winding, hilly, rural roads
has frequently been confronted with the problem of passing a slower vehicle
ahead and has either driven many laborious miles waiting for an opportune
time to pass or has ventured doubtfully into the passing maneuver on the chance
that it could be accomplished without mishap. If the motorist had sufficient
information about conditions on the highway ahead—whether there is an on-
coming vehicle in the opposite lane and whether there is enough room on the
highway to pass the car ahead and clear the oncoming vehicle—the passing
maneuver not only could be executed more safely, but the volume of the traffic
served by the roadway would be increased by minimizing inherent delays caused
by slower vehicles.
The Public Roads research and development program has turned to electronics
in the search for a method of providing information that the driver needs to
pass vehicles safely on 2-lane highways. Results of experiments conducted on a
2-lane roadway with an elementary passing aid system, PAS I, are described
in this article. The purpose of these experiments was to determine whether
drivers would rely on information supplied electronically to indicate the absence
of opposing vehicles when visual sight distance was limited. Encouraging results
of the experiments, as shown by acceptance of electronically indicated passing
opportunities, have prompted the planning of more advanced experiments and
the development of a more sophisticated passing aid system. Work is now under-
way on Passing Aid System II (PAS II), which is expected to be installed on 15
miles of 2-lane rural highway during 1969.
PUBLIC ROADS ® Vol. 35, No. 3
+z | Aen :
ae
Bye de OFFICESOE
RESEARCH AND DEVELOPMENT
BUREAU OF PUBLIC ROADS
Introduction
RIVING on high-volume, winding, 2-lane
rural highways is a problem that is known
to most motorists. Restricted sight distances,
oncoming traffic, and adverse environmental
conditions make it difficult or impossible for
a motorist to pass slow vehicles, and any one
of these conditions not only encourages unsafe
passing attempts but also tends to decrease
average vehicle speed. Furthermore, difficult
passing situations, such as those existing on
winding mountain roads and in dense traffic,
discourage all passing attempts and encourage
the unsafe practice of tailgating.
Less known to the layman is the decrease
in the capacity of a highway caused by the
inability of motorists to pass vehicles ahead.
ven when passing sight distances are ade-
quate, traffic volume still may reach only
30-70 percent of the roadway’s capacity
(volume/capacity ratio). Unfortunately, as
most 2-lane rural highways do not have
unrestricted passing sight distances, the
volume/capacity ratio is further reduced, and
61
INITIATION OF
PASSING
MANEUVERS
1ST
FLAG
2ND
FLAG
BEGIN TEST RUN
AT STATION [80
FAIRBANK HIGHWAY
RESEARCH LABOR-
ATORY
Figure 1.—Test route and operations for Passing Aid System, I.
the effect is to reduce the number of passing
opportunities and create more traffic inter-
ferences—slowdowns, accidents, ete.—which
reduce the service volume.
Even after the Interstate System has been
completed, more vehicle-miles will be traveled
on rural highways than on rural sections of
the Interstate System. From this fact alone,
it is evident that rural highways must be
made safer. More than one-third of all
accidents on these highways at present are
rear-end collisions. Head-on collisions do not
oecur as frequently as rear-end collisions—
about one-fifth of the accidents are head-on
collisions—but they are likely to be more
severe. Both types of accidents, however,
involve the interaction of two or more drivers
and their vehicles.
According to past research, a driver cannot
estimate, with any degree of precision, the
absolute speed of a vehicle ahead or the rate
at which his own vehicle is approaching it
until the two vehicles are only a few hundred
feet apart. Also, according to past research,
when the two vehicles are this close to each
other, there is not enough time for the driver
to modify his speed, especially if he is traveling
at a speed typical of those on the highways
today.
To avoid rear-end collisions, drivers need
to be given reliable information about the
speed of the vehicle ahead, or about relative
speed or closure rate. Speed patterns of pairs
of vehicles involved in rear-end collisions
support the fact that the driver of the colliding
vehicle lacked information on the vehicle
ahead—more than one-third of the passenger
cars were traveling at speed differences greater
than 30 m.p.h. prior to collision. In normal
traffic, however, less than 1 percent of pairs
of cars travel at speed differences exceeding
30 m.p.h. Head-on collisions, although oc-
62
PASSING CAR
SYSTEM
CONTPOL UNIT
TRAFFIC
SIGHT RESTRICTORS DETECTORS
INSTALLED FOR
MEST
LEAD CAR
curring less frequently than rear-end collisions,
must be given equal attention because of
their severity.
Research has shown that the average driver
requires approximately 9 seconds to initiate
and complete a passing maneuver on a 2-lane
rural highway. Thus, if one vehicle traveling
at 70 miles per hour overtakes another, a
9-second passing maneuver requires that the
highway ahead be clear for a distance of
more than 1,800 feet. At this distance, not
only are drivers unable to estimate the relative
speed of a vehicle in the opposite lane but
they are incapable of determining whether
that vehicle is stopped, in fact, whether its
motion is toward them or away from them.
Many 2-lane, bidirectional rural highway
sections are without sight distances of 1,800
feet and, accordingly, are marked to prohibit
passing. Moreover, the degree of precision
in executing the passing maneuver has become
increasingly important as traffic volumes have
increased and vehicles in the opposite lane
are being encountered more frequently.
Public Roads, through its national program
of highway research, is endeavoring to increase
travel safety on rural highways by developing
methods to give the driver adequate environ-
ment information on 2-lane roadways. This
information may relate to speed, acceleration,
closure rate, or other information about both
the vehicle ahead and the vehicle the
opposite lane.
in
The objective of this research is to develop
a system to aid drivers solve discrimination,
judgment, information, and vehicle control
problems on 2-lane rural highways, and,
consequently, raise highway service volumes
and increase traffic safety.
Applications of electronics technology are
being explored as a means to aid drivers in
making judgments and
during overtaking
WEST ENTRANCE TO
FAIRBANK HIGHWAY
RESEARCH STATION
OPPOSING CAR
IN OPPOSING LANE
*
STA.NO. | q
SIGHT RESTRICTORS
INSTALLED FOR TEST
passing maneuvers. A specific application
the development of an electronic aid syste
that will provide the driver with informatic
as to the presence, location, and speed |
vehicles in the opposing lane. It was post)
lated that over a specified distance of sufficie)
length, considering all combinations of vehie
velocities, road grades, ete., a go or no-go tyj
of system could be employed.
A full-seale mockup of an electronic passil
aid system has been constructed and test
on the 2-lane access road to the Public Roa
Fairbank Highway Research Station
McLean, Va. Summarized in this article a
the concepts, experimental tests, and pr
cedures used to determine whether drive
can and will use this electronic aid syste
known as Passing Aid System I, or PAS
The willingness of drivers to use PAS I, al
their ability to apply it successfully as an a
in passing vehicles on 2-lane highways, pl
vides an indication of the advisability
developing a more advanced passing @
system. .
Considerations in Developing a
Passing Aid System
To speed development of passing aid §
tems, tentative decisions were made:
Drivers would be given distance informati
and possibly speed information or time-
meeting information, and (2) the system I
to be compatible with existing operations
that drivers of vehicles unequipped Ww
electronic hardware could continue to use
highway.
The following basie questions need to
answered before a_ fuil-scale passing
system can be made available for public 1
e Will drivers pass if they have informat
about the absence of opposing vehit
within a critical distance?
August 1968 ® PUBLIC RO.
Table 1.—Test subjects used as vehicle drivers
Distance driven during last 12
ar months
Driving
\ Phase Test Age experience Occupation
subject 2-lane City
rural Freeways streets
highways
Numober Years Years Miles Miles Miles
20 3 UU CON taeenee eee a 75 20 5, 500
2 20 3 SUNGeD ssser eae aoe 100 1, 500 5, 500
3 21 5 Sitdenitewns ees 500 2, 500 2, 500
4 ee 5 Studentees.2 ae 4, 000 2, 000 2, 500
5 23 7 BPR Meee ee Se 500 2, 000 2, 000
Preliminary Tests__-- 6 27 10 BPR Bngineer!--_.2- | 4, 000 7, 000 5, 000
i 22 6 Ntudenpsss eS 100 800 2, 000
8 30+ 30 pecretary == oe 8, 000 3, 000 1, 000
9 24 8 Secretary..............| 3,000 6, 000 3, 000
10 62 15 PeCralar vers ce easy 500 500 1, 000
11 23 7 IB Din Mees Ae eB || 500. 2, 000 2, 000
12 29 13 BPR Engineer______- 14, 000 6, 000 10, 000
13 22 5 MOCTOLAT Yess eee ee 5, 000 10, 000 5, 000
Experiment 1___-_..- 14 21 1 STIGSR Taree oan ee 2, 500 5, 000 2,500
15 20 3 Studentesseaq e vee | 1, 000 2, 000 2, 000
16 58 40 BPR Foreigner ______- 800 1, 700 38, 000
yee Le FY Eee eee, OL ewes piled! seal inoteghetac yb. ce, Howe ce are wy eM |anney, og Bie. guy | ag
18 35 19 BPR Engineer_____-.- 4, 000 4, 000 4, 000
19 31 10 BPR Engineer___-_-_-_-_- 3, 000 1, 000 6, 000
20 45 18 BPR Hngineer_._____- 4, 000 8, 000 1, 000
Experiment 2----.--- 21 26 11 BER Engineer 2. > | 8, 000 8, 000 2, 000
22 18 2 Studentias ewes eee eee es 38, 000 1, 000
23 51 30 Spiritualist___ yt Oo 38, 500 7, 000 3, 500
24 35 13 Hlousewites. es. eee | 4, 000 8, 000 3, 000
25 20 5 Stident# ee Se | 5, 000 1, 000 6, 000
What criteria should be employed in deter-
mining the critical distance at which an
opposing vehicle is brought to the attention
of the passing driver?
Will drivers employ other distance informa-
tion about opposing vehicles in addition
to the critical distance?
/ Will drivers use opposing vehicle speed
information in making a passing maneuver?
Will drivers use time-to-meeting informa-
tion in making a passing maneuver?
How long does it take drivers to adapt to
the new system?
| What instructions should be given to drivers
to make it easy for them to learn how to use
the new system?
What criteria should be employed in deter-
mining how far apart the vehicle detectors
shall be?
Are there any side effects and reliability
considerations that may affect the operation
'/ of the new system—driveways, cross roads,
| | Opposing cars passing other opposing cars,
, |steep gradients, stopped vehicles, etc.?
'How will environmental conditions—rain,
snow, ice, darkness—affect system opera-
tions and how can they be overcome?
What will be the costs and the benefits of a
passing aid system?
Should information be given to drivers in
visual, auditory, or tactile form?
Preliminary answers to some of these ques-
ons were obtained from experimental work
‘ith the PAS I system reported here. How-
ver, most of the questions will be answered
uring PAS IT operations.
Objectives of PAS I Study
The initial objectives of the first experi-
1ents with PAS I were as follows:
To determine whether drivers, even though
their sight distance is restricted, will pass
when they are informed that there are no
Opposing vehicles within a specific distance.
|
q UBLIC ROADS ® Vol. 35, No. 3
e To ascertain that drivers will use speed
information about the opposing vehicle to
aid them in passing.
¢ To obtain an indication of how long it takes
drivers to adapt to a new system with one
set of instructions.
e To determine whether clearance distances
between passing and opposing vehicles at
the end of passing maneuvers are adequate,
based on use of 1,300-foot signal distances.
The results of experiment 1 indicated that
drivers will make selective use of passing aid
information given to them by electronic means.
Additional planning of more sophisticated
passing aid systems is now well underway.
Description of PAS I
The first experiment was based on the use
of a mockup of the Passing Aid System. The
mockup, called PAS I, was installed on the
access road to the Fairbank Highway Re-
search Station and covered a distance of
approximately 0.7 mile. A simplified sketch of
the PAS I test setup is shown in figure 1.
The west bound direction road was used for
the passing maneuver in which one vehicle, the
passing car, was to overtake and pass another
vehicle, the lead car, according to coded
messages issued by the electronic passing aid
system. The eastbound lane was used as the
opposing lane in which an oncoming vehicle,
the opposing car, approached the two west-
bound vehicles in the east lane to provide a
situation that required the driver of the passing
vehicle to execute the passing maneuver in
time to avoid a collision or to stay in his lane
behind the lead car. Sight restrictors, installed
along the roadway, obstructed the
view of the road ahead and simulated the blind
condition on 2-lane rural highways caused by
hills and curves. Traffie detectors were spaced
44 feet apart in the lane used by opposing
vehicles, and as the opposing vehicle moved
driver’s
over each detector, an intermittent audible
signal The
signal, which could be received by the passing
was. transmitted. intermittent
car, was detectable at any point within 1,300
feet ahead of the opposing vehicle.
Four conditions could exist for the driver
of the passing car: (1) No signal—the system
was not operating, (2) a steady uninterrupted
signal—the opposing lane was clear of traffic
for at least 1,300 feet, (3) the beginning of
an intermittent signal—there was a moving
vehicle 1,300 feet ahead the
lane, and (4) repetition of the intermittent
1,300
in opposing
signal—a moving vehicle was within
feet ahead in the opposing lane. The fre-
quency/second of the
increased with the speed of the
vehicle. After the beginning of the signal,
the number of intermittent signals and the
intermittent
signals
opposing
speed of the opposing vehicle indicated the
clearance distance between the two vehicles.
Figure 2.—Diagrammatic representation of variables.
63
Test Subjects and Vehicles
Test subjects used in the two experiments
were obtained from the student body of
George Washington University, the Bureau
of Public Roads staff and the general public.
Information about the drivers is shown in
table 1.
The passing and opposing vehicles driven
by the test subjects were 1967 4-door sedans—
Dodge, Valiant, and Plymouth—with the
following specifications: automatic transmis-
sion, power steering, power brakes, 6 cylinder,
225-cu. in. cylinder displacement, and 145-
brake horsepower.
Table 2.—Minimum passing-sight-distance
for design of 2-lane highways !
2 | Assumed Minimum
Design speed | passing passing-sight-
| speed distance
m.p.h, | m.p.h feet
30 | 30 800
40 } 40 1, 300
50 | 48 1, 700
60 55 2,000
70 | 60 2, 300
' Source: Blue Book, Geometric Design Rural Highways
1959, p. 211, : et
The lead car used in experiment 1 was a
1966 4-door Ford sedan with automatic trans-
mission. The lead car in experiment 2 was a
1967 4-door Mercury sedan with automatic
transmission, power steering, power brakes,
200-brake horsepower, 8 cylinders, and cyl-
inder displacement of 289 cu. in. In general,
the test drivers considered the power and
performance of the vehicles they drove to be
adequate.
The combination of the driver and the
vehicle he drove for the first time presented
significant variables that were significant in
determining the acceptance of a passing aid
system.
Description PAS I Study Variables
The variables considered in the preliminary
studies are itemized in the following list, and
where applicable, they are shown in figure 2:
Distance
D =distance between passing and oppos-
ing car.
D.=signal range generated ahead of op-
posing car, 1,300 feet.
D;=D where passing may begin—any-
where between the two flags.
D,=D when passing maneuver begins.
D.=D where passing maneuver ends.
Speed
Vi=speed of lead ear.
V.=speed of car following lead car prior
to passing maneuver.
V3=speed of opposing car.
Time
T,=time required to pass.
T;=time for test car to reach juxtaposi-
tion with opposing car after having
completed passing maneuver.
64
Time, in addition to distance and specd,
was observed in the hope that it would serve
as a check—distance=speed X time—and be
useful for the period covered by the passing
maneuver when acceleration and _ speeds
change significantly.
Two types of test runs were used in the
experiments—radio and control. In the radio,
or PAS I, test runs, the electronic passing aid
system was used by the driver of the passing
car to overtake and pass the lead car. The
control test runs were made without the use of
the passing aid system and were included in
the experiments to provide a basis of com-
parison in analyzing the effectiveness of
PAS I. The 2-lane test roadway had a design
speed of 70 m.p.h., a posted speed limit of 30
m.p.h., and two long, 3-degree curves with
a tangent between them. The posted speed
limit was not in effect for the test runs.
Discussion of Variables
Preliminary test data, collected prior to
experiments 1 and 2, indicated that several
variables would have to be controlled.
The first variable was sight distance. Sight
distances were so large it would have been
difficult to determine whether there was any
difference in the frequency of passing maneu-
vers between the control and simulated PAS
I test runs. To decrease passing opportunity
and more closely simulate driving conditions
that would exist on a rural mountainous road,
temporary panels were installed (see fig. 1) to
restrict the sight distance. To insure com-
parable passing opportunities for the test and
control situations, the sharpest curve, near
the midpoint of the test road, was used. The
part of the curve between stations 150 and
125 was selected as the section of roadway
PASSING AID SYSTEM I
Test Subject:
EXPERIMENT NO, 1
Observer:
where passing maneuvers could begin. He
sight distances were 620-1,300 feet, bi
temporary panels reduced them to 400-5
feet. The sight distance was based on ft
ability of the driver in the right lane to s
any part of a vehicle in the opposing la
The second variable to be controlled wa
the frequency at which an opposing vehiel
was encountered in the passing area. For th
control phase of preliminary test runs, driver
were instructed to drive the way they nor
mally drive, but the frequency of passin
maneuvers seemed abnormally high. Tes
drivers confirmed this by volunteering th
information that normally, on the open high
way, they would not pass if the sight distane
were comparable. 4
Three possible reasons were considered fe
the incongruity between drivers’ statement
and actions. The first was that the test roa
was always cleared of other traffic so the
passing manuevers could be based solely o
the position and speed of the opposing te
vehicle. This was a definite requirement fe
study of PAS I and consequently, to perm
a comparison, it was also a requirement f¢
the control phase. Because test drivers kne
there would be only one vehicle in the opposir
lane and that the driver of the opposing ¢&
would be aware of the passing maneuver, the
were more willing to pass. They apparent
believed that they were not fully responsib
for the passing maneuver and its possible co
sequences, as they are on the open road. —
Another possible reason for the discrepant
between the statements and actions of ft)
drivers was that they were speaking in gener
terms based on normal operating speeds. F
example, table 2 gives minimum passing sig
distances of 800 and 1,300 feet at speeds
FIELD DAT
Date: Time Begin:
Station Number
Time-0,01 min,
Passing | Opposing car (Units=passing car. Tens=opposing car, Twenties=passed car) ks zero
Run car Start t :
No. |Series |(m.p.h.) | m.p.h.| Station Lyi Rati eo Se M93 ive Remarl
51 [149 us | 1os| roe | 11 132 PASS
h 102 Sift
= 30 | Pass
ree
es a
—
bate eee
Patines
ea ia
PF
C
Q
| sees
= a
pe] * | 4s |
he | =| [oe
~
=
fe} ~
ss
fo | = | 30 [4s
Figure 3.—Sample data sheet, experiment No. 1.
, a oe Ue _—
August 1968 ® PUBLIC RO
sa
0 and 40 m.p.h. respectively. If these speeds
ere considered normal, it would have been
nsafe to pass after installation of the tem-
orary sight-restrictor panels, because the
aximum sight distance available in the des-
nated passing area was less than 550 feet.
t 20 m.p.h., however, the minimum passing
ight distance would have been approximately
e same as the sight distance available, and
e test subjects should have been willing to
ass with or without the use of PAS I. The
referred approach was to study conditions
which passing maneuvers normally were
ot feasible, and it was decided to eliminate
st runs based on a lead car speed of 20 m.p.h.
A third possible reason for the incongruity
vetween drivers’ statements and actions was
at the opposing vehicle and the passing
‘ehicle seldom were near the passing area
}limultaneously, and drivers may have realized
hat it was usually safe to pass the lead car.
Any of these possibilities or combinations
f them, could have accounted for the high
vassing frequency in the control phase of the
reliminary tests. To eliminate the first
|vossibility, decreased driver responsibility,
e following driving instructions were issued
“lo the test subjects; these instructions re-
laced the game aura of the experiment with
ne of responsibility:
|| Entering the car.—‘Please fasten your safety
“helt. The purpose of this research study is to
“\Inalyze how you drive so we may develop
ids to other drivers.”’
Test runs, control phase.—‘‘Please start the
iar. Drive as you normally would on this
-lane highway. Follow the car ahead. There
4 vill be traffic coming toward you in the
| |Pposing lane. If in your judgment you would
‘ vormally pass the car ahead, you are free to do
0 by beginning your passing maneuver some-
vhere between the two red flags along the left
ide of the road. If you do not consider it safe
0 pass, continue to follow the car ahead.
a Drive safely. Take no chances. Drive in a
aanner similar to the way you drive on the
q \ighway. Any questions?”
| Test runs, passing aid phase.-—‘‘When you
i \ear a continuous tone, from your radio
~ eceiver, the opposing lane is clear of moving
ASS ch °
~~ traffic for at least 144 mile. When a vehicle
“3 moving toward you in the opposing lane
— loser than 14 mile, you will hear beeps on
— he radio. If you desire to pass, you may use
— |he radio signals to aid you in deciding whether
_ ir not to pass.
— “As before, if you do choose to pass the car
— head, the passing maneuver should start in
—\he_ area between the red flags. Any
_ {uestions?”’
—| The second possibility was eliminated by
_ lisearding the 20-m.p.h. test runs, mentioned
_\arlier, and the third by increasing, for each
Vest subject, the percentage of runs in which
here was no opportunity to pass in the
~ yassing area. For the no-passing situation to
~»eceur in the designated passing area, it was
~ tecessary to specify not only the lead car and
~ »pposing vehicle speeds, but also the stations
“irom which the vehicles would begin each
jest run.
: JUBLIC ROADS © Vol. 35, No. 3
START
M2
STATION
POSITION,
m
t
t= 0 e
FINISH CONTACT
. €
ELAPSED TIME FROM t,, HUNDREDTHS OF MINUTES (0.0i)
Figure 4.—Vehicle positions at which station numbers and elapsed time was recorded.
Experiments and Precedure
Data were collected in two series of tests—
experiment 1 and experiment 2. In both
experiments the test subjects were used in
pairs. For a test run, one subject would operate
the passing vehicle and the other subject the
opposing car. For the next test run, the drivers
exchanged assignments so that each driver was
used coming and going in each pair of runs.
Experiment 1
Data for experiment 1 were recorded on the
form shown in figure 3. The first five vertical
colums at the left contain the previously dis-
cussed control variables. The run number
indicates the individual trips on the test road
during which a passing maneuver could occur.
The column originally indicated the trip
sequence for both drivers, but midway through
experiment 1, this arrangement was deter-
mined to be undesirable, as one driver of each
pair of drivers would operate the opposing
vehicle in one run, then operate the passing
vehicle in the next run under identical test
conditions. Consequently, he could recall the
starting position of the opposing vehicle, its
speed, the clearance distance available for
passing, or any one of these factors, to formu-
late a predetermined pass or no-pass decision.
In the field it was decided to eliminate the
65
Speed combination
Ne. PEA eee, om
Table 3.—Data summary for experiment 1, test subjects Nos. 12-17 !
Begin run
Lead car
| Opposing
car
Lead car
Opposing |
car
Opposing lane clear for more than 1,300 ft. at
Ist flag
Pass/No-pass frequency
mS (abet be Seated Ta’ Le take ye een. _
cP We +? . F we .
Passing percentage
Control
PASI
Control
Pass
m.p.h.
30
m.p.h.
15
180
180
Station No.
180
180
180
180
180
180
180
180
180
180
1 Test subjects are listed in table 1.
Speed combination
Station No.
27
50
88
27
50
88
27
50
88
27
50
88
Number
No-pass
Pass
No-pass
Number
Number
Number
Test
Percent
Total
Percent
J
PASI
Test
Percent
4
4
Total ;
fi
Percent
Table 4.—Data summary for experiment 2, test subjects Nos. 18-23 !
Begin run
Lead car
| Opposing
car
Lead car
Opposing
car
Opposing lane clear for more than 1,300 ft. at
Ist flag
m.p.h.
30
m.p.h.
15
Station No.
180
180
180
180
180
180
180
180
1 Test subjects are listed in table 1.
Speed combination
Station No.
517
ai
88
Pass/No-pass frequency
Passing percentage
Control
PAS I
Control
Pass
Number
2
No-pass
Number
4
Pass
Number
4
No-pass
Number
9
2
Test
Percent
Total
Percent
Table 5.—Data summary for experiments 1 and 2 combined, test subjects Nos. 12-17 and 18-23 !
Begin run
1 Tests subjects are listed in table 1.
'
Pass/No-pass frequency
Passing percentage
PASI
Test
Percent
66
Total
Percent
Opposing lane clear for more than 1,300 ft. at
y 7 Ist flag Control PASI Control PAS I
Lead car | Opposing | Lead car | Opposing
car car
Pass No-pass Pass No-pass Test Total Test Total
m.p.h. m.p.h. | Station No. | Station No. Number | Number | Number | Number | Percent | Percent | Percent | Percen
30 15 { 180 27, Mies 2. 500 ees en) Cee ene. Oe ire + 6 5 2 40:45 spheres 71 --
‘ | 180 88 pe ke | Sie eee see iN. 4 9 5 9 31” Av eeeat 36 ae |
‘Total: passing fee se eee l= Percents si" a2! pe | Pe nee ee ae SD Ca eieeee Sree 48
: 180 27 ViO8. 25. Sst nn ees oo ee SRS eg 2S 3 us 17 4 30
30 a Ba. A No kee eas = Reine eH i 5 25 || i ae eee
| ‘Total pussitre® ons eee percent} eine Cac Ne eee 2 eee 20 ol ae “68
: Be ad 180 27 BME 2 Se ee eS a 0 6 0 6 0
45 a rt Bab | Nose eee ae CaP es & aa 1 13 1 1 7 | ae ieee ae
‘Dotal Passive) oe Beach te POTCRNE 22} ee 2a Sos eas NS Be ee eee ee (at Se a
45 45 { 3 27 OOK 226s ae Ta Bae on, bs eee Nw Oe 3 9 8 20 aa | ee 50 in
August 1968 ® PUBLIC RO
i ‘ y
ibility of predetermined decisions by a
ndom selection of each succeeding test run
om those remaining to be made. Accordingly,
e data in experiment 1 were considered to be
ee of predetermined pass or no-pass decisions.
In the second column of figure 3, control,
eans without the use of the passing aid sys-
2m, and radio, means with use of the passing
d system. The data under pass. car are the
eeds of the lead car and passing car prior to
carting the passing maneuver. The speed of
1e opposing car and the station from which it
“}varted are given in the next two columns for
ach test run. In all test runs, the lead car
nd the test vehicle started from a stationary
~osition at station 180. The results of pre-
minary test runs, before PAS I was op-
rational, were the basis for determining the
eselected variables.
The instrumentation for the experimental
jrocedure was simple. Distances were deter-
ined by relating vehicle positions to the elec-
onic detector stations numbered consecu-
vely from the west end of the test road;
rations were 22 feet apart. Speeds for the
‘ad car and the opposing car were preselected,
nd the drivers accelerated to the constant
needs and maintained them by referring to
—n,e speedometer in the vehicles. Two-way
idios were used to communicate among the
— ree vehicles and disseminate the following
formation: road clear, meaning the route is
‘ear for the test run; flag, meaning the begin-
™ jing of the length of roadway where a passing
iq (neuver could be initiated; start, meaning
_ ie driver has started the passing maneuver;
, |nd finish, meaning the driver has completed
le passing maneuver and has returned to the
ght lane. The relations between vehicle posi-
ons and elapsed times are shown in figure 4.
. ‘he term contact, shown at time t,, means that
le passing vehicle and the opposing vehicle
re at the same position on the roadway after
_. ompletion of the passing maneuver.
| In figure 3, m;, m:, m3, and my are the
> ation ieiors of the test vehicle at flag,
__ ‘art, finish, and contact respectively; my, ™M42,
‘nd mj are the station locations of the oppos-
_ 1g car when the passing car signals flag, start,
nd finish; and m3 is the recorded position
f the lead car at the finish signal. In the time
olumns of figure 3, ts, ts, and t, are the
lapsed times in hundredths of a minute (0.01)
—om flag to start, finish, and contact, respec-
ively.
taf
otal
| (xperiment 2
Experiment 2 was basically a continuation
f experiment 1, but was different in two
ar, were provided for the drivers, (2) The
;)umber and distribution of runs, were adjusted
ccording to control variables so that there
“vould be matching control and radio (PAS 1)
IS.
The predetermined orders of runs for experi-
aent 2, which were based on the theory of
it jandom numbers, are shown in figures 5 and 6.
hese two orders, A series and B series,
UBLIC ROADS ® Vol. 35, No.
aie
a ea al) hale ee Eg
> a ‘ ve fi.
oes aad a =|? " > 4
PASSING AID SYSTEM I
Test Subject:
EXPERIMENT NO, 2
FIELD DATA
|
Observer: Date: Time Begin:
Lead & Station Number : Time-0,01 ae pe
passing | Opposing car (Units=passing car, Tens=opposing car. Twenties=passed car) t,= zero
A car Start i
Series |(m.p.h,) | m.p.h.| Station] ™1 Lay! ape bat | ee ee
| 1 [conrees Ys 1s- 88 [ vw T#
—_+} +
a a RE #5 | 47
Meee so.) vs a7
=H + +
‘ a5 (SR 02.7, |
——~ —+ — + T —
RADIO “45- tog a7 |
” ma
6 30 1s~ 27
T t
7| " SOSH S A 2 | |
oS Fowlers) Wes
2 ae a |
: “5 | 4s- | 99 fe
+
11 Ys Yo ae
12 4S | (SF | 8F
+ 4
dete
14
24 | go Bia wane
4s See | ee, |
45 a7 rate
a | |
eee oe ws AR bw x [po eles RS IES |
at
5 es = ih [ :
20 A el a
Figure 5.—Sample data sheet, experiment No. 2,
eliminated possible bias that could have oc-
curred if the test runs had been selected in the
field, and decreased the possibility that drivers
would know the position of the opposing car.
Test runs in A series and B series were similar,
only the sequence was different.
The percentage of test runs that were pass-
ing maneuvers in the control and PAS I phases
are shown, in figures 8, 9, and 10 respectively
for experiment 1, experiment 2, and experi-
ments 1 and 2 combined. Data points on the
Series A.
graphs are the percentage subtotals shown on
tables 3, 4, and 5. For example, in table 6 and
figure 8 it is shown that for a lead car speed
of 30 m.p.h., an opposing car speed of 15 m.p.h.,
and three opposing car starting stations, 59
percent of the control test runs were passing
maneuvers. To make valid comparisons of
passing percentages between the control and
PAS I phases, the proportions of the runs
assigned to the different opposing car starting
PASSING AID SYSTEM I
Test Subject:
stations would have to be equal for both
_¥ETIELD DATA]
EXPERIMENT NO, 2 FIELD DATA
Observer: Date: Time Begin:
Lead & Station Number Time-0,01 min,
passing | Opposing car (Units=passing car. Tens=opposing car, Twenties=passed car) t= zero
Run B car Start
No. |Series |(m.p.h,) | m.p.h,| Station
1 |conTROy “45— 4S | a7
2 ies our | (e288
2 ae 30 Her) SRS
Bak 30 [sien
5 |RADIO| 70 ¢5 | 8&
“4 JO Tre 8s
+ GON Ao at 7
‘é 30 is | #7
> AS (Ss | 27
Petiefes |o isenlir se
is eT Le \ leet
27
/-
YS 1§-
4s SS-
ee ee ee ee ed
ola |JNYN ID} onl] e®l_ wl d]trio;}o;oaolinr
Figure 6.—Sample data sheet, experiment No. 2, Series B,
LEAD CAR SPEED,MPH.
OPPOSING CAR SPEED, MPH.
Figure 7.—Estimated position of passing car at which signal from opposing car would first be received.
phases. This requirement was met for experi-
ment 2 (fig. 9) but was not for the other
experiments, which may explain the possible
misalinement of the control phase curves of
figure 8. The data in all three figures indicate
the apparent increased percentage of test runs
having passing maneuvers for transitions from
the control phase to the PAS I phase.
Statistical Tests
Chi square tests and confidence limit
intervals were used to determine the statis-
tical significance of the results. An advantage
of the chi square test is the yes or no answer
obtained. However, in situations where data
do not meet minimum requirements, the chi
square test is not applicable. Confidence limit
bands can be based on any size sample, but
conclusions can vary with interpretation of
the bands. Both approaches were used with
emphasis given to the one considered most
applicable to the particular analysis being
made.
Chi square tests
The chi square statistic takes into account
the similarities of samples that occur by
chance alone, regardless of whether the
samples are from the same or different popu-
lations. A calculated value of the chi square,
based on observed data, can, be compared to
standard tabulated values of chi square shown
in textbooks on statistics (1).1 Depending on
the percentage level of confidence desired, the
comparison can infer whether any difference
in two samples is likely to have occurred by
chance alone. If an existing difference did not
occur by chance alone, then the difference is
significant.
The chi square tests used in this report
were based on the use of 2 X 2 tables (1
degree of freedom), and a tabulated value of
chi square equal to 3.84 for the 95-percent
confidence level. The 95-percent confidence
1 The italic numbers in parentheses identify the references
listed on p. 76.
2 Statistical tests that were made for each analysis have
been assembled and are available from the Office of Research
and Development, Bureau of Public Roads, %, Managing
Editor, Public Roads Magazine.
68
P45) 45
OPPOSING CAR STARTING STATION 88 88
140
level is commonly used and accepted in
research.
Confidence intervals
To estimate the mean of a population, it is
helpful to have not only a sample mean but
also a measure of the margin of error of the
sample mean. A way to do this is to specify a
zone, based on the sample mean, within which
the population mean, lies. This zone is called a
confidence interval, and the end points of
this interval are called confidence limits. The
probability that the interval will include the
population mean is stated as a percentage and
is referred to as the confidence level. The
95-percent confidence level was used in the
research reported here.
The control phase of the experiments was
the population, or real world, used as a basis
of comparison. Because the control phase was
100
CONTROL PHASE
(WITHOUT USE OF Pas)
VARIABLES.
60
OPPOSING—
40 CAR SPEED
15 MPH.
PROPORTION OF DRIVERS PASSING, PCT.
OPPOSING—
CAR SPEED
45 MPH,
pa koe
Figure 8.—Percentage of drivers passing—with and without PAS—experiment 1.
GRAPHS AT LEFT ARE NOT DIRECTLY
COMPARABLE TO THOSE AT RIGHT AS
TEST RUNS FOR EACH PHASE WERE
BASED ON DIFFERENT CONTROL
LEAD CAR SPE
also a sample, the test basically was a com
parison of two sample intervals. If the range
of the two confidence iatervals were general)
similar, the samples were from the sam
population. If the ranges of the confidene
intervals were generally different, then th
samples were from different populations.
Confidence limits for each proportion wer
obtained from an Ordnance Engineering De
sign Handbook (2). The upper and lower con
fidence limits were obtained from tables fe
samples of fewer than 30 observations an
from graphs requiring interpolation for sam
ples of more than 30 observations.
The results of the statistical tests have bee
assembled as yes or no answers to question}
given in table 6.2 For example, if a statistics}
test determined that a slight increase in pas:
ing frequency was insignificant, the answer i
PAS I PHASE
(WITH USE OF PAS)
I
OPPOSING-
CAR SPEED
FAS?
ou
ED, MPH.
August 1968 © PUBLIC RO/
owes 55 2 ee’ aa ay ee
for this analysis, the use of confidence limits on
experiment 2, indicated that PAS I did not
increase passing percentage.
Analysis 4
The analysis, based solely on the use of
PAS I indicated that when lead ear speed was
increased from 380 to 45 m.p.h., there was a
decrease in passing frequency. The statistical
tests based on experiment 1 produced answers
to the contrary, or answers with doubtful
conclusions because of the unbalanced sample
distribution. Experiment 2, with balanced
sampling, produced the most reliable con-
clusions which were supported by the conclu-
sions from the combined data of experiments |
and 2. The accepted conclusion is reasonable,
considering the fact that as traffic speed
increased, fewer passing maneuvers were re-
quired to maintain desired speed.
Analysis 5
When signaled clearance distance at the first
passing opportunity was more than 1,300 feet,
the analysis of PAS I indicated an increase in
passing frequency with the use of PAS I,
when compared to the control phase. The
statistical tests were in agreement for each of
Figure 9.—Percentage of drivers passing—with and without PAS—experiment 2. the experiments.
.ftable 6 and in the following discussion would
state that there was no increase in passing
af frequency.
Analyses of Statistical Data
“A summary of the analyses of the primary
data, frequency of passing maneuvers, is given
| pin table 6. The analysis number at the left is
followed by the question that the analysis
poses. Answers to the question, based on use
of the chi square test and confidence intervals
for each experiment, are indicated in the
olumns at the right.
Experiment 2 was the only experiment that
proportionate distribution of test runs
ith regard to the control variables for each
sample. Comparisons of data for experiment
2 were therefore favored over those for experi-
ment 1 and 1 and 2 combined.
Analysis 1
_ The first analysis was made to determine
whether PAS I, compared to the control phase,
increased the percentage of passing. The com-
een for each experiment was based on all
ata. Each statistical test applied to the differ-
ont experiments indicated that use of PAS I
did increase the percentage frequency of
ssing maneuvers.
alysis 2
The analysis of PAS I, compared to the
trol phase for lead car speeds of 30 m.p.h.,
nd all the statistical tests used, indicated
iat PAS I increased the passing percentage.
nalysis 3
C ompared to the control phase for lead tar
speeds of 45 m.p.h., the analysis of PAS I
failed to show conclusively that it increased
J
iS Sa
PY BI LIC ROADS ® Vol. 35, No. 3
fae
|
on he
7 :
a
oo)
passing percentage. Chi square tests were
Analysis 6
When signaled clearance distance was less
than 1,300 feet, the analysis of PAS I compared
to the control phase indicated no change in
passing frequency. Statistical tests were in
agreement for all experiments, though it should
PAS I increased passing percentage for experi- he noted that data were below the minimum
ment 1 and 1 and 2 combined, both of which required for chi square tests in experiments |
were unkalanced samples. The one reliable test and 2.
limited because of sample distribution and/or
the minimum data criteria for the test. The
use of confidence intervals indicated that
PAS I PHASE
(WITH USE OF PAS)
AT LEFT ARE NOT DIRECTLY
TO THOSE AT RIGHT AS
TEST RUNS FOR EACH PHASE WERE
BASED oN DIFFERENT oN
_ OPPOSING-CAR
SPEED 45 M.PH.
Figure 10.—Percentage of drivers passing—with and without PAS—experiments 2 and 3
combined.
Table 6.—Results of statistical tests
Analysis No., description, and question
Answers for each experiment
Chi square test
Confidence interval
land 21
combined
11
Did PAS I, compared to control phase, show larger per-
centage of passing?
Did PAS I, compared to control phase for lead car speed of
30 m.p.h., show larger passing percentage?
Did PAS i compared to control phase for lead car speed of
Did passing fuk cedunaes decrease with use of PAS I and
increase in Jead car speed (30-45 m.p.h.)?_....._------------
Did PAS I, compared to control phase, increase passing
Did PAS i compared to conte phase, give similar passing
percentage ‘when signal clearance was less than 1,300 ft.?_-_-
Did PAS I, compared to control phase show increase in
land 21
combined
passing for:
a Lead car speed of 30 m.p.h., opposing car speed of 15
m.p.h., opposing car starting station 27
b Lead car speed of 30 m.p.h., opposing car speed of 15
m.p.h., opposing car starting station 88 3
c Lead car speed of 30 m.p.h., opposing car speed of 45
m.p.h., opposing car starting station 27
Lead car speed of 30 m.p.h., opposing car speed of 45
m.p.h., opposing car starting station 88 3
Lead car speed of 45 m.p.h., opposing car speed of 15
m.p.h., opposing car starting station 27
Lead car speed of 45 m.p.h., opposing car speed of 15
m.p.h., opposing car starting station 8& 3
Lead car speed of 45 m.p.h., opposing car speed of 45
m.p.h., opposing car starting station 27
Lead car speed of 45 m.p.h., opposing car speed of 45
m.p.h., opposing car starting station 88 3
1 Conclusions weakened by sample distributions.
2 Does not satisfy criteria for chi square test.
Table
3 Signaled clearance distance is less than 1,300 ft. when
the passing vehicle reaches the permitted passing area.
7.—Clearance distance between passing and opposing vehicles at completion of
passing maneuver, experiments | and 2 combined
|
Control variables Test runs
favo ; Average
Speed of | Beginning Phase With clearance
Speed of opposing | station for Total With passes distance
lead car car opposing passes and
car data
m.p.h. m.p.h. number : number number number Jeet
is 4 ie 27 { : Ceol Naa tetzel ceuNe es ee ze 4 : - ne
re eo 8 OE RS See wee EEE eae Se ee a eee tee eee )
i & ie i (Controlca> a a hee es 14 5 3 }
30 i : | : : 180
a0 = (PASI Ee 14 5 5 374
30 5 97 DB LGtep ty ho} Ee Aa eae 4 3 300
Re EE eee ee a get a su
0 5 88 VA Clee WMT LED MOiSe Raa é 2 0
3 45 NP AS. TG. 9 Grol teams 8 2 1 242
15 15 97 || CONGEOLS 2 a CRY eee 8 0 OFA gt eee
Ween ee De kee Me Be pole ae ees if 0 Oude the PASS
rs FE 19 ONtrOle ee 284s see ee 14 1 it 154
ay) Pee | \PAS ep. eee 13 1 1 132
3 an oF i Con Urol meq nee ee eee 13 4 3 96
mre a5 a (PAB TS eee en ears eet: 8 7 798
45 45 88 {Control CR a SNS fae ey Ses 10 0 Og) a eee
J [LE JAS tee Ose eS es 8 1 Of Ae ee,
Analysis 7
Tn this analysis the confidence intervals were
developed for the control and PAS I phases of
eight combinations of lead car speed, opposing
car speed, and opposing car starting station.
The question for each combination was ‘‘does
the use of PAS I increase the frequency of
passing maneuvers when compared to the con-
trol phase?’ Results of statistical tests can be
categorized as follows:
(1) Those combinations based on the oppos-
ing car starting from station 88 showed no
70
increase in passing maneuvers. The signaled
clearance distance at the first passing oppor-
tunity for each combination was less than
1,300 feet. The findings were in agreement
with those of analysis 6.
(2) Those combinations, based on the
opposing car starting from station 27, showed
an increase in passing maneuvers when the
lead car speed was 30 m.p.h. and no increase
in passing maneuvers when the lead car
speed was 45 m.p.h. The findings were in
agreement with those findings of
2 and 3, respectively.
analyses
Ly
ew cn «tema Lh, Sah FPP RS Bes Hy
ee ee |
Clearance at End of Passing
Maneuvers
One of the objectives of the experimen
was to determine whether clearance distane
between the passing and opposing vehicl
at the end of the -passing maneuver, basec¢
on use of the 1,300-foot clearance distance
according to data from combined experiments)
1 and 2, are shown in table 7. Data in the
table are classified at the left by the control
variables: lead car speed, opposing car speed,
and opposing car starting station. For both
the control and PAS I phases each classifi
cation shows the total number of test run
observed, number of test runs that wer
passing maneuvers, number of test runs with
passes for which clearance data were obtained,
and average clearance distance between the}
passing and opposing vehicle at the end
the completed passing manuever.
The use of PAS I when compared to the
control phase, based on the lead car speed
of 30 m.p.h., increased the clearance distance
at the end of the passing maneuvers. For
lead car speeds of 45 m.p.h., the clearance
distance decreased.
Summary
A full-scale mockup of a Passing Ai
System (PAS I) was installed and tested on}
a short section of 2-lane highway. Result}
of a limited experiment showed that wher}
sight was restricted to a distance considerably
below the 800 feet of passing sight distanes
required for a design speed of 30 m.p.h.
drivers made selective use of passing-distane
information given to them electronically. A
operating speeds of 30 m.p.h., drivers mad
significant use of PAS I when ee distance
exceeding 1,300 feet were indicated elee
tronically, and the passing percentage wa
substantially increased (see fig. 9). At 4
m.p.h., drivers used PAS I less frequently
The sight distance for passing at 45 m.p.f
is approximately 1,500 feet, but the presem
design of PAS I provided only 1,300 feet fo
passing.
The one set of instructions used in thes
experiments introduced drivers to the passin
aid system but did not adapt them to 1
Use of the passing aid system is required. Th
experience gained was too limited to cor
clude that the drivers had satisfactorily adap
ed to the system.
Based on use of PAS I, clearance distane¢
between the passing vehicle and opposir
vehicle at the end of the passing maneuyi
were adequate when lead car speeds wel
30 m.p.h., but inadequate when lead ¢
speeds were 45 m.p.h. The control phase wi
used as the basis for this conclusion.
Results of the study were favorable for tl
development and use of passing aid system
Although the range of conditions under whi¢
such systems would be useful may not be :
broad as anticipated, further testing wit
passing aid systems is justified.
(Continued on p. 76)
August 1968 © PUBLIC ROA
ee
Al
Deviation from the mean travel speed on
the Interstate System increases the proba-
bility of involvement in an accident.
\nterstate system Accident Research
‘Study Il, Interim Report II
BY THE OFFICE OF
RESEARCH AND DEVELOPMENT
UREAU OF PUBLIC ROADS
, Introduction
HE RESULTS of an analysis of the
“il effects that speed variance among vehicles,
| evel of enforcement, and interchanges have
a accident and involvement rates are pre-
ented in this report—the second interim re-
‘ vort on data collected for the Interstate
)iystem Accident Research Study II. The
) bjectives of the research and related study
“ rocedures were described in Interim Report
| (1).1 The data used in the analysis presented
ere were collected by 20 State highway
y partments (see fig. 1).
if 4
»| Speed Variance Among Vehicles
We
_| It has been shown in past research that the
everity of a given accident will increase as
lhe speed of the vehicles prior to collision
fereases (2, 3). However, the chance of
| ing involved in an accident, at least on
_ = and 4-lane main rural highways, having no
| eperel of access (2), has been shown to be
elated to speed variance, or deviation from
j
1 The italic numbers in parentheses identify the references
ted on p. 75.
Wh
Q
UBLIC ROADS ® Vol. 35, No. 3
the speed of the traffic stream. It was sought
in the analysis reported here to determine
whether speed variance contributes to acci-
dent involvement on the Interstate System
as well. Only accidents occurring between
9 a.m. and 4 p.m. on mainline units were
used in the analysis to correspond with the
speed data and vehicle classification data
collected for the same period. For this study,
a mainline study unit is defined as any section
of the Interstate highway that is not more than
10,000 feet in length and homogeneous
throughout, with respect to its geometric
characteristics. Speed data were not obtained
during the hours from 4 p.m. to 9 a.m. Speed
change lanes, although classified as separate
units, were included in the category of main-
line units. Ramps, crossroad units, front-
age roads, and other units were not included
in the analysis. To further reduce the number
of variables, 2-lane two-direction mainline
study units were eliminated from the analysis;
however, both urban and rural sections were
studied but were not separated.
In determining the effect of speed variance—
not used here in the statistical sense—only
rear-end and angle collisions and same-direc-
tion sideswipe accidents, occurring between
Reported by JULIE ANNA CIRILLO,
Mathematician,
Traffic Systems Division
9 am. and 4 p.m., were considered. The
assumption was that the effect of vehicular
speed differences could best be determined by
accidents involving two or more vehicles
traveling in the same direction; thus head-on,
single vehicle, and pedestrian accidents were
not included. Speed data were submitted by
the States on EAM (Electronic Accounting
Machines) ecards in the format shown in figure
2. The coded information
percentage of traffic traveling in each speed
group. The data were not gdjusted in any
manner but were used precisely as submitted
by the States.
The mean travel speed for each study unit
was obtained by accumulating the products
of the midspeed for each of the speed group-
ings—for example, 45 m.p.h. for the speed
group 40-49 m.p.h.—and the percentage of
the vehicles traveling within the speed group,
then dividing the final total by 100. The mid-
speed used for the wnder-40-m.p.h. speed group
was 37.5 m.p.h. for rural areas and 32.5 m.p.h.
for uiban areas. The midspeeds used for the
80-m.p.k.-or-more speed group was 85 m.p.h.
These midspeeds were determined from speed
trend data collected on Interstate highways
in many States (4).
represented the
rg:
STATES THAT SUBMIT TED
DATA USED IN ANALYSIS
[at 7) STATES NOT PARTICIPATING
we
OTHER PARTICIPATING STATES
Figure 1.—States participating in Interstate System Accident Research, Study I.
Results of Speed Analysis
Results of the analysis indicated that a
reduction in the variation of speed among
vehicles should significantly reduce accidents.
The procedure for determining involvement
rates, as related to mean speed, was similar to
that reported by Solomon for 2-lane and 4-lane
rural highways (2). Involvement rate is the
number of involvements per 100 million vehi-
cle-miles and implies a vehicle involved in an
accident. Thus, one accident involving three
vehicles is counted as three involvements. The
curve shown on figure 3 was plotted on the
basis of variation from the mean speed of each
unit. The involvement rate at each speed, for
each study unit, was related to the variation
from the mean speed of the study unit. For
each accident that occurred on a study unit
being used, the speed of each vehicle involved
in the accident was subtracted from the mean
speed of the study unit. For example, suppose
the mean speed of a study unit was computed
to be 60 m.p.h. and a vehicle involved in an
accident on this unit was traveling at 55
m.p.h.,? then this involvement would be re-
2 Speeds submitted by the State and probably extracted
from accident report forms.
72
ported as having occurred at a variation of
minus 5 m.p.h. from the mean speed. All such
data were grouped together to obtain a data
point; results of these calculations have been
summarized and are shown in table 1. The
data points weer plotted on figure 3, in addition
to points obtained by Solomon. As these are
daytime data, only Solomon’s daytime curve
was plotted to provide a common basis for
comparison.
In table 1, it is shown that the lowest
involvement rate occurred at approximately
+12 m.p.h. above the mean speed of a study
unit. One might expect the lowest involvement
rate to occur at the mean speed; but the
variation inherent in collecting and estimating
speed data is possibly the reason that the
lowest involvement rate occurs at +12 m.p.h.
above the the
magnitude of the variation increased, either
above or below the mean speed, the involve-
mean speed. However, as
ment rate increased. These results
remarkably similar to those reported by
Solomon. This curve is shifted slightly to the
right of Solomon’s curve (see fig. 3), in which
the lowest involvement rate occurred at ap-
proximately +8 m.p.h. for daytime accidents;
were
but Solomon’s study was conducted on 2-lé
and 4-lane main rural highways that had |
control of access. Usually, on this type
conventional highway, the average speed)
lower than on the Interstate highway; 1
mean speed was about 52 m.p.h. on conv
tional rural highways and 59 m.p.h. on Int
state highways.
Table 1.—Involvement rate by deviati
from mean speed
Deviation from mean | Involve- | Vehicle- Involve
speed ments Iniles |mentrat
m.p.h. Number | Millions | Numbe
—30.0 to —34.9 82 sis 63, 222
— 256.0 to —29.9 129 1,93 6, 673
—20.0 to —24.9 109 14. 03 re
—15.0 to —19.9 245 86. 86 282
—10.0 to —14.9 259 180. 91 143
—5.0to —9.9 356 519. 52 68
0.0to —4.9 321 756. 41 42
+0.1to +4.9 290 772. 84 37
+5.0to +9.9 162 566. 95 28
+10. 0 to +14.9 46 180. 38 25
+15. 0 to +19. 9 21 60. 13 35
+20. 0 to +24. 9 14 10, 29 136
+25. 0 to +29.9 10 8. 25. 307
+-30. 0 to +34. 9 13 wed 11, 627
1Involvement rate=number of involvements per
million vyehicle-miles.
August 1968 © PUBLIC ROA
“
ine re ee 6 0, strrecstel swat 1677 3h Ty00 Results shown in table 2 indicate that no
TE : trend can easily be established between an
789 0 1 2.13 4 1 16 17 16 19 70 21 20 23 24 2 a7 28 29.30 3! sa 3: r iner ase i rarni 3, arrests bf ice atr
To qo oo ON do[ejo 000 AloTo 0-0 O]ui0 00 ool 0 0 O1d'0 Da cles 00 D]O}0 00 OlOjo NT q ners ee teeta) npias pone nae
‘COLUMNS 60-69--PERCENTAGE TO THE NEAREST PERCENT OF VEHICLES and the mean speed of travel or the involve-
114 1/" TRAVELING IN EACH SPEED GROUP EXAMPLE: 8 PERCENT O8 PUNCHED IN Two- '!! ment rate on a study section. Further investi-
2222), COLUMN FIELD OR 8 IN ONE-COLUMN FIELD. gation of enforcement related to traffic volume
syaabssasle: COLUMN NO. DATA(SPEED GROUP) 4313 and other variables will be undertaken in the
| 60 UNDER 40 MILES PER HOUR (ONE COL, FIELD) future.
444gedaaadiad, 6l-62 40-49 MILES PER HOUR 44/4
Bece Rees pee ae Beis 5 51s5585.5.555! Effect of interchanges on accident rates
epi 467-68 70-79 MILES PER HOUR =A se In the analysis of the effect of interchanges
69 BOMILES PER HOUR OR MORE (ONE COL. FIELO)**/{ ep te a he
) naia7 eee rre een t 1771/7 CUES ACE RL ee etal eay bet accident rates, all units were divided into
urban or rural sections. Eac ainline i
| 8828191809 8.6:8 8 88/66 BB 88/8 0 8 bic Go 8 Beis 08 Bigg OG 5/3 6 8 eS closastlessetes c alelee sella see aasslclosects : ; ct as ae solace i
| | | ae | was then positioned by its proximity to an
| q aig a 9 3 919 a8 9 18 aye 998) A annie a <7 ee ari ears
|egesspasyseessieasestlassseeaagslcoasseegsupessascesseensuesgsceesdessgezseeeics2e8| interchange. Because each unit was located
‘a JBM 5282 IBM ” SERVICE ‘BUREAU
soe between two interchanges, ahead and behind,
accidents were assigned to the nearest inter-
change. Units equidistant from two inter-
changes were divided between the two
interchanges.
Figure 2.—Speed data collected for Study II.
Of more importance is the generally lower 100,000 p
scident rate observed on Interstate highways.
Ithough the average speed of vehicles on the
iterstate System is 7 m.p.h. higher than on
ynventional main rural roads, the Interstate
‘stem can better accommodate differences
_ vehicle speeds, with the exception of very
ow-moving vehicles. It appears, however,
iat with respect to accident involvement on
PaReys, as well as on conventional rural
ghways, both very high and very low speeds
be dangerous, and it is differences in speed
nong vehicles that cause hazardous situa- 10,000
ons. The hazard of slow-moving vehicles on
gh-speed highways is indicated by the sharp
se in involvement rate for vehicles traveling
)m.p.h. below the mean speed.
evel of law enforcement
An attempt was made to investigate the
feet of the level of law enforcement on mean
eed and accident involvement. Data sub-
iitted represented the average number of
. gal warnings, written warnings, arrests, and
, plice patrol hours per mile per year on the
. iterstate System. Only study sections for
, hich this information was provided were
, ted in this analysis. Law enforcement
iformation was requested for mainline study
uits only, but in several States, this informa-
bn was not available and these sections were
~ used in this portion of the analysis.
, |Enforcement data were collected on an
erage daily basis, and speed data were
_ (lected for daytime hours only. Therefore, 8
was assumed that the average daily enforce- 100 --—
ent data were proportional to the level of
- ytime enforcement. As speed data were
(lected for daytime hours alone, only those
cidents—single and multivehicle—occurring
‘tween 9 a.m. and 4 p.m. were used. The other
iteria for the data base in this analysis were
@€ same as the ciiteria used in the speed
alysis—that is, no distinction was made
1000
_ SOLOMON'S CURVE
~ CONVENTIONAL
RURAL HIGHWAYS
INVOLVEMENT RATE (PER 100 MILLION VEHICLE MILES)
INTERSTATE —
HIGHWAYS _
‘htween urban or rural sections; 2-lane two- -30 —20 -l0 O +10 +20 +30
} ection mainline units were eliminated; only VARIATION FROM MEAN SPEED, MPH.
,lainline units and speed-change lanes were
ed, and only traffic volumes between 9 and Figure 3.—Accident involvement rate by variation from mean speed on study units.
p.m. were considered.
P| BLIC ROADS ® Vol. 35, No. 3 4.
0.2 MILE
122
Distances were measured from the midpoint
of each study unit to the gore (beginning of the
ramp) and were recorded in discrete codes
which represented continuous intervals of
unequal length. The accident rates—the num-
ber of accidents per 106 million vehicle-miles—
for both between-interchange units and at-
interchange units were calculated. An inter-
change was assumed to extend from the
beginning of the deceleration lane taper to
the end of the acceleration lane taper. Thus,
the following interchange units were included
in this analysis:
Deceleration lanes including taper
Acceleration lanes including taper
Exit ramps
Entrance ramps
Mainline units between speed-change lanes
Combined acceleration-deceleration lanes
The at-interchange accident rate, shown on
figure 4, was a weighted combination of the
accident rates for each of these units. Accident
rates were not calculated for crossroad units,
terminal areas between ramps and crossroads,
frontage roads, and local streets.
In interpreting the results of the analysis, it
is essential to note that the only variables
considered were the distances between the
study unit and the interchange, and the
classification of the section—rural or urban.
No other variables were considered.
74
Figure 4.—Accident rate by type of interchange unit.
Table 2.—Involvement rate by type and level of enforcement
Type of enforcement
Level of enforcement
per mile per year
Mean
speed
Vehicle miles
Involvement rate
-. &
Cars
Trucks
Oral warnings?_____-
Written warnings--_--
Car hours of police
patrol.
ATT AS(S > Were. aon
1 Involvement rate=
2? No data available.
Number
hessithan's222-eee
6-14 er ee a
15-342 Seek he ee re
150-290. 2B ca
300-549 Se ee
C001. 199 Fe ae
Over 1,200 ?_
Unknown_.-_-----
160-2992 ee
300-549. ee
GO0=1 2190) re
Overt 2004232 a2
IN KT10 Wilke ee oe
4 Pes SO ee BA 2
710-149. eee
150-2002. ee ee
S00=549 - 2 SiS 22
600-1199 Soi
Over] 2005 2= eee
Unk Gwithoseee ee
600-1,199_ _
Over 1,200 ?
Unknown._..-..-
number of involvements per 100 million vehicle-miles.
Involvements
Cars Trucks
Number | Number
960 124
140 59
636 125
537 103
458 54
308 39
74 9
14 Z
Cee ieee ye
337
731
296
210
460
79
Millions
1, 856
287
769
632
398
354
42
August 1968 © PUBLIC ROA
Millions
371
71
173
124
68
49
Cars
Number
52
49
Trucks
Number
33.
82
ale
EXIT SIDE
Distance to exit-ramp Accidents
Accident
nose ahead
rate !
‘
ENTRANCE SIDE
Distance to entrance-ramp
‘ Accidents
nose behind
Accident
rate 1
URBAN
URBAN
Number
messmmans -2 miles... 2... _. 722
a> i 1, 209
b- . i 786
1.0-1.9 miles 280
Eaaro THUHCS..-- en a acne 166
4.0-7.9 miles
More than 8 miles 3
Number
131
: Number
Less than .2 miles 426
.2- .4 miles 1, 156
.5- .9 miles 655
1.0-1.9 miles 278
2.0-3.9 miles 151
4.0-7.9 miles
More than 8 miles 3
Number
122
125
RURAL
amos ihan 9.2 miles: .-- . 2= = ..2
Lah
5- .
1.0-1.9 miles
2.0-3.9 miles
4.0-7.9 miles
More than 8 miles 3
1 Number of accidents per 100 vehicle-miles.
2 Small sample size.
3 No data available.
Results
‘The results reported below indicate that in
ban areas, proximity of a study unit to an
terchange had a substantial effect on the
ecident rate. A similar effect, of less magni-
ide, was observed in rural areas for study
nits near entrance ramps.
retween-interchange accident rates
As shown in table 3, the accident rate de-
reased on urban sections as the study unit
yas positioned farther away from an exit
amp; the highest rate occurred in units less
han 0.2 mile from the exit ramp. This de-
ease was substantial to a distance of approxi-
\}aately 2 miles from the ramp. Also, as a unit
vas stationed farther from the entrance ramp
i) |rea, the accident rate decreased, although
ot uniformly. Moreover, the rates on both
ides of the interchange were fairly compara-
Je. On rural sections, however, the change in
ates, as a unit was positioned closer to the
) aterchange, was not significantly altered; and
| the exit direction it remained almost con-
tant. Thus, in urban areas proximity to inter-
i hanges seemed clearly to affect the accident
Interchange unit
Less than
i
1.0-1.9 miles
2.0-3.9 miles
4.0-7.9 miles
More than 8 miles 3
Table 4.—Interchange-mileage relations by
area type
ATCA GN DOs: ek eees or eer Urban Rural
Number of interchanges_____ 718 942
Number of miles___________- 1, 380 3,919
Interchanges per mile ______- 02 . 29
Miles between interchanges _ iL) 3.4
rate, probably because in urban areas inter-
changes occur almost twice as frequently as in
rural areas (table 4), and usually carry much
higher volumes.
At-interchange accident rates
Accident rates are presented, in figure 4, for
each type of at-interchange unit; sample size is
indicated in table 5. The total at-interchange
accident rate was, as noted above, a weighted
combination of the accident rates for each
separate unit type computed for the 100 mil-
lion vehicle-mile base.
When interpreting the figure, it is important
to note that only exit ramps and entrance
ramps are shown. Included in these calcula-
tions are ramps which are part of diamond-
type interchanges, outer connections and loops
Table 5.—Accident rate by interchange unit and area type
Area type
Rural
Urban
Vehicle-
miles
100 Million
2.51
0. 57
6. 52
Deceleration lane
Exit ramp
Area between speed change lanes
Entrance ramp 0. 59
Acceleration lane 3. 68
0.49
Accidents
Number
Accident
rate !
Number
186
Vehicle- | Accidents
miles
100 Million
5, 83
199 346 1,48
554 85 11. 87
95 1. 61
8. 40
Accident
rate 1
Number
137
Number
1, 089
546
1, 982
348
370
1, 159
1, 461
2,45
14. 36
31. 64 6, 792
1 Accident rate = number of accidents per 100 million vehicle-miles.
UBLIC ROADS ® Vol. 35, No. 3
of cloverleaf interchanges, semidirect and
direct connections, and slip ramps.
The accident rate for urban interchanges is
substantially higher than for rural inter-
changes, as these areas carry more traffic,
making merging and diverging maneuvers
more difficult. Because of higher right-of-way
and construction costs, urban interchanges
tend to be less standard in design, are more
complex, and are confined to smaller areas
than rural interchanges. These factors increase
conflict possibilities, and also make entrance
and exit maneuvers more difficult. The excep-
tionally high accident rate on entrance ramps
in urban areas may be caused by inadequate
acceleration lanes, or the lack of them, on
many sections, necessitating vehicles to stop
at the bottom of the ramp before moving into
the traffic stream. Also, the unavailability of
sufficient gaps in the main traffic stream makes
it difficult to merge into moving traffic.
The accident rate on the mainline decreases
after the deceleration lane has been passed
(figure 4). It appears that after the decision
point at the deceleration lane has been passed,
the chances of an accident decrease.
From this brief analysis it can be determined
that sections in proximity of interchanges
experience a higher accident rate than other
sections. Ramps have much higher accident
rates than speed-change lanes (and paralleling
main roadway) and these, in turn, have
generally higher rates than the other portions
of the main roadway.
Conclusion
The results reported demonstrate that on
the Interstate System, as the speed of a
vehicle varies from the mean speed of traffic,
either above or below the mean speed, the
chance of the vehicle being involved in an
accident increases; that the level of enforce-
ment has little or no apparent effect on the
mean speed or on the accident experience of
a study section; and that proximity to inter-
changes, especially in urban areas, appears to
affect significantly the accident experience of
the study section.
Although these results are not conclusive
they provide some insight into areas in which
more intensive research should be conducted,
such as interchange spacing and utilization and
more effective methods of speed control.
REFERENCES
(1) Interstate System Accident Research—
Study I1—Interim Report I, by Julie Anna
Cirillo, Highway Research Record 188, Traf-
fic Accident Research, 1967, pp. 1-7.
(2) Accidents on Main Rural Highways Re-
lated to Speed, Driver and Vehicle, by David
Solomon, U.S. Government Printing Office,
Washington, D.C., July 1964.
(3) The Lone Driver Involved in Injury-
Producing Accidents, by William E. Schotz,
Cornell Aeronautical Laboratory Inc., Cor-
nell University, Buffalo, N.Y., Apr. 25, 1966.
(4) Traffic Speed Trends, U.S. Department
of Commerce, Bureau of Public Roads, March
1965.
75
Standard Plans for Highway Bridges, vol. I1,
Structural Steel Superstructures, 1968 ($1.00 a
copy), is a revision of the 1962 edition with
respect to bridge widths and current design
specifications. These plans are intended to
serve as a useful guide in the development of
suitable and economical bridge designs. An
effort has been made to give sufficient infor-
mation on all plans so that they may be readily
modified in the preparation of contract
drawings.
Passing Aid System I, Initial Experiments
ACKNOWLEDGMENTS
The cooperation of many persons of the
Public Roads staff was necessary for the test
driving and the collection and analysis of
data for the experiments of Passing Aid
System I. The persons assisting with the
collection of field data included Raymond
Greenwell, Santo Salvatore, Howard §. Ellis,
and John Porter. The electronics system,
76
NEW PUBLICATION
The volume contains plans for simple span
I-beam and welded girder superstructures
from 20 feet to 180 feet, simple span two-
girder with floor system superstructures from
120 feet to 200 feet, and continuous 3-span
I-beam and welded girder superstructures
with center spans from 50 feet to 240 feet.
Bridge roadway widths used are 28 feet with
H15-44 live load for low traffic volume low
(Continued from p. 70)
designed by Raleigh Emery, was kept opera-
tional during driver test runs by Messrs.
Novean and Porter. Margaret Cormack
assisted with the scheduling of test subjects,
the administrative details of ordering equip-
ment, and the preparation of the final report.
Statistical aid was supplied by Dr. Harry
Weingarten and Mrs. Phyllis Mattison. Mr.
David Solomon provided overall guidance and
optimistic enthusiasm in conducting the
analysis of Passing Aid System I.
design speed roadways, 44 feet with HS20-44
live load for the standard 2-lane two-diree.
tional roadway and 40 feet with HS20—44 live
load for the standard 2-lane one-directiona
roadway.
One series of simple span I-beam super
structures with Interstate loading and F
variable width roadway is included in th’
plans.
REFERENCES
(1) Introduction to Statistical Analysis, }
Wilfred J. Dixon and Frank J. Massey, J)
McGraw Hill Book Co., Ine., First Editio}
1951, pp. 184 and 185.
(2) Experimental by Maj
tibbons Natrella, U.S. Dept. of Commert
National Bureau of Standards, 1963, p. T-é)
Statistics,
August 1968 ® PUBLIC RO/S
a
*
|
A list of the more important articles in PuBLIc Roaps and title
heets for volumes 24-34 are available upon request addressed to
ureau of Public Roads, Federal Highway Administration, U.S.
epartment of Transportation, Washington, D.C. 20591.
The following publications are sold by the Superintendent of
ocuments, Government Printing Office, Washington, D.C. 20402.
rders should be sent direct to the Superintendent of Documents.
repayment is required.
ecidents on Main Rural Highways—Related to Speed, Driver,
and Vehicle (1964). 35 cents.
ggregate Gradation for Highways: Simplification, Standardiza-
tion, and Uniform Application, and A New Graphical Evalua-
tion Chart (1962). 25 cents.
erica’s Lifelines—Federal Aid for Highways (1966). 20 cents.
nnual Reports of the Bureau of Public Roads:
1963, 35 cents. 1964, 35 cents. 1965, 40 cents. 1966, 75 cents.
1966 supplement, 25 cents.
(Other years are now out of print.)
e Bridge to Your Success—Opportunities for Young Engineers
in the Bureau of Public Roads (1967). 30 cents.
alibrating and Testing a Gravity Model for Any Size Urban
Area (1965). $1.00.
apacity Analysis Techniques for Design of Signalized Intersec-
tions (Reprint of August and October 1967 issues of PUBLIC
ROADS, a Journal of Highway Research). 45 cents.
Construction Safety Requirements, Federal Highway Projects
(1966). 55 cents.
onstruction Safety Requirements, Federal Highway Projects
(1968). 50 cents.
orrugated Metal Pipe Culverts (1966). 25 cents.
Creating, Organizing, & Reporting Highway Needs Studies
(Highway Planning Technical Report No.1) (1963). 15 cents.
Federal-Aid Highway Map (42 x 65 inches) (1965). $1.50.
Federal Laws, Regulations, and Other Material Relating to High-
ways (1965). $1.50.
ederal Role in Highway Safety, House Document No. 93, 86th
Cong., Ist sess. (1959). 60 cents.
Freeways to Urban Development, A new concept for joint
development (1966). 15 cents.
! Guidelines for Trip Generation Analysis (1967). 65 cents.
Highway Beautification Program. Senate Document No. 6, 90th
Cong., Ist sess. (1967). 25 cents.
ighway Cost Allocation Study: Supplementary Report, House
Document No. 124, 89th Cong., 1st sess. (1965). $1.00.
Highway Finance 1921-62 (a statistical review by the Office
of Planning, Highway Statistics Division) (1964). 15 cents.
ighway Planning Map Manual (1963). $1.00.
Highway Planning Technical Reports—Creating, Organizing, and
Reporting Highway Needs Studies (1964). 15 cents.
Highway Research and Development Studies, Using Federal-Aid
Research and Planning Funds (1967). $1.00.
ighway Statistics (published annually since 1945) :
1965, $1.00, 1966, $1.25.
(Other years out of print.)
ighway Statistics, Summary to 1965 (1967). $1.25.
Highway Transportation Criteria in Zoning Law and Police
Power and Planning Controls for Arterial Streets (1960). 35
cents.
Highways to Beauty (1966). 20 cents.
Highways and Economic and Social Changes (1964). $1.25.
Hydraulic Engineering Circulars:
No. 5—Hydraulie Charts for the Selection of Higway Cul-
verts (1965). 45 cents.
q PUBLICATIONS of the Bureau of Public Roads
No. 10—Capacity Charts for the Hydraulic Design of High-
way Culverts (1965). 65 cents.
No. 11—Use of Riprap for Bank Protection (1967). 40 cents.
Hydraulic Design Series:
No. 2—Peak Rates of Runoff From Small Watersheds (1961).
30 cents.
No. 3—Design Charts for Open-Channel Flow (1961). 70
cents.
No, 4—Design of Roadside Drainage Channels (1965). 40
cents.
Identification of Rock Types (revised edition, 1960). 20 cents.
Request from Bureau of Public Roads. Appendix, 70 cents.
The 1965 Interstate System Cost Estimate, House Document No.
42, 89th Cong., 1st sess. (1965). 20 cents.
Interstate System Route Log and Finder List (1963). 10 cents.
Labor Compliance Manual for Direct Federal and Federal-Aid
Construction, 2d ed. (1965). $1.75.
Amendment No. 1 to above (1966), $1.00.
Landslide Investigations (1961). 30 cents.
Manual for Highway Severance Damage Studies (1961). $1.00.
Manual on Uniform Traffic Control Devices for Streets and High-
ways (1961). $2.00.
Part V only of above—Traffic Controls for Highway Construc-
tion and Maintenance Operations (1961). 25 cents.
Maximum Desirable Dimensions and Weights of Vehicles Oper-
ated on the Federal-Aid Systems, House Document No. 354,
88th Cong. 2d sess. (1964). 45 cents.
Modal Split—Documentation of Nine Methods for Estimating
Transit Usage (1966). 70 cents.
National Driver Register. A State Driver Records Exchange
Service (1965). 20 cents.
Overtaking and Passing on Two-Lane Rural Highways—a Litera-
ture Review (1967). 20 cents.
Presplitting, A Controlled Blasting Technique for Rock Cuts
(1966). 30 cents.
Proposed Program for Scenic Roads & Parkways (prepared for
the President’s Council on Recreation and Natural Beauty),
1966. $2.75.
Reinforced Concrete Bridge Members—Ultimate Design (1966).
35 cents.
Reinforced Concrete Pipe Culverts—Criteria for Structural De-
sign and Installation (1963). 30 cents.
Road-User and Property Taxes on Selected Motor Vehicles
(1964). 45 cents.
Role of Economic Studies in Urban Transportation Planning
(1965). 45 cents. °
Standard Alphabets for Highway Signs (1966). 30 cents.
Standard Land Use Coding Manual (1965). 50 cents.
Standard Plans for Highway Bridges:
Vol. II—Structural Steel Superstructures (1968). $1.00.
Vol. 1V—Typical Continuous Bridges (1962). $1.00.
Vol. V—Typical Pedestrian Bridges (1962). $1.75.
Standard Traffic Control Signs Chart (as defined in the Manual
on Uniform Traffic Control Devices for Streets and Highways)
22 x 34, 20 cents—100 for $15.00. 11 x 17, 10 cents—100 for
$5.00.
Traffic Assignment Manual (1964). $1.50.
(1963). 15 cents.
Traffic Safety Services, Directory of National Organizations
Transition Curves for Highways (1940). $1.75.
Typical Plans for Retaining Walls (1967). 45 cents.
UNITED STATES POSTAGE AND FEES PAID ©
U.S. GOVERNMENT PRINTING OFFICE
GOVERNMENT PRINTING OFFICE
DIVISION OF PUBLIC DOCUMENTS
WASHINGTON, D.C. 20402
OFFICIAL BUSINESS
r
If you do not desire to continue to receive
this publication, please CHECK HERE (J;
tear off this label and return it to the above
address. Your name will then be removed
promptly from the appropriate mailing list.
I
\
RA
DOT LIE
Iti
MA
loses Meal
Deedee a
“Salat erw ee et «
Se id
Se
Con
Ee dy neal hee ae i
Peace Nene cig ata PU EO we
Dene e ee cae add a ane
ad sree e ~*
wre ers)
Spree eee
Re ae Se
fe di i te Be
are ore
Petar tah a eid
PS a os Oe
Peers
ra
ee Ne Se od on oe
ee
Sian 5 aed ioe he fap are ae
SPP cain mE Woe SP nie oh
Ae Lee ore wom ey
ESE ert eye
vires
NEP I s,
re eed sak
ee ee eae se
ese ake Mae
2
~s Ves
ote ow ey
are ae
oat adad aaeatek Me
= # Cer ;
eel ae haa be is dh ae ie Ad
i he a POET See ~
Pek os Cy) ee Sl ee ae re
APA PPO PE PPL EO
os 6 oe Te ee = Se enw
ee SO pee se Fa we Ba 1 er
oe te ad OP A oh eee ee ee ll at i SEI > Ray
ee ee OO EE EE PEG ROE OE a
tor hdr tn te Mele
Live Music Archive
Librivox Free Audio
Metropolitan Museum
Cleveland Museum of Art
Internet Arcade
Console Living Room
Open Library
American Libraries
TV News
Understanding 9/11